Retroviruses undergo frequent genetic alterations which in many instances may contribute to the pathogenicity of the viruses. Murine retroviruses, for example, readily undergo recombination with endogenous sequences of mice, leading to host range variants. Such variants have been directly implicated in the activation of cellular oncogenic genes. Variants which escape the host immune response or which exhibit an altered pathogenicity may be the result of point mutation of lentiviruses, such as the equine infectious anemia virus (EIAV) and the human immunodeficiency virus (HIV). This project is focussed on the occurrence and mechanisms of genetic alteration in retroviruses as well as the consequences of such alterations. In previous work we have utilized a number of monoclonal antibodies (mAbs) to identify and isolate recombinant host range variants from mice which were inoculated with murine leukemia viruses (MuLVs), or from mice which harbor endogenous MuLVs and spontaneously generate such variants. We have now identified three antigenic classes of the recombinant viruses which encompass all of the isolates examined. Structural analyses of the recombinants suggests that the three antigenic classes correspond to three distinct classes of endogenous gene sequences. Furthermore, we have found that different MuLVs recombine with different sets of endogenous sequences to preferentially generate distinct classes of viruses. Utilizing in vitro-constructed chimeric MuLVs, we found that the selectivity for a particular class of recombinant maps to the same regions of the genome previously found to influence the tissue-specific replication of the inoculated viruses. In preliminary experiments we have demonstrated pseudotyping of the inoculated virus resulting in an altered in vitro host range. Thus, it is possible that the generation of a particular recombinant influences the sites of replication of the inoculated virus. Alteratively, replication in different tissues may influence the class of recombinant viruses generated. Another aspect of this project has focussed on the determination of the point mutation rate of retroviruses. Previous estimates of the point mutation rate of retroviruses have not been carefully controlled for the number of replication cycles and/or the target size for mutation. By strictly controlling the number of replication cycles and directly examining the virion RNA genome, we have determined that the mutation rate for a murine retrovirus is approximately 20-fold slower than previous estimates reported in the literature.